ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/AnyEvent/lib/AnyEvent.pm

Comparing AnyEvent/lib/AnyEvent.pm (file contents):
Revision 1.44 by root, Mon Apr 7 19:42:18 2008 UTC vs.
Revision 1.165 by root, Tue Jul 8 23:07:26 2008 UTC

1	=head1 NAME	1	=head1 NAME
2		2
3	AnyEvent - provide framework for multiple event loops	3	AnyEvent - provide framework for multiple event loops
4		4
5	Event, Coro, Glib, Tk, Perl - various supported event loops	5	EV, Event, Glib, Tk, Perl, Event::Lib, Qt, POE - various supported event loops
6		6
7	=head1 SYNOPSIS	7	=head1 SYNOPSIS
8		8
9	use AnyEvent;	9	use AnyEvent;
10		10
…		…
14		14
15	my $w = AnyEvent->timer (after => $seconds, cb => sub {	15	my $w = AnyEvent->timer (after => $seconds, cb => sub {
16	...	16	...
17	});	17	});
18		18
19	my $w = AnyEvent->condvar; # stores wether a condition was flagged	19	my $w = AnyEvent->condvar; # stores whether a condition was flagged
		20	$w->send; # wake up current and all future recv's
20	$w->wait; # enters "main loop" till $condvar gets ->broadcast	21	$w->recv; # enters "main loop" till $condvar gets ->send
21	$w->broadcast; # wake up current and all future wait's	22
		23	=head1 INTRODUCTION/TUTORIAL
		24
		25	This manpage is mainly a reference manual. If you are interested
		26	in a tutorial or some gentle introduction, have a look at the
		27	L<AnyEvent::Intro> manpage.
22		28
23	=head1 WHY YOU SHOULD USE THIS MODULE (OR NOT)	29	=head1 WHY YOU SHOULD USE THIS MODULE (OR NOT)
24		30
25	Glib, POE, IO::Async, Event... CPAN offers event models by the dozen	31	Glib, POE, IO::Async, Event... CPAN offers event models by the dozen
26	nowadays. So what is different about AnyEvent?	32	nowadays. So what is different about AnyEvent?
…		…
29	policy> and AnyEvent is I<small and efficient>.	35	policy> and AnyEvent is I<small and efficient>.
30		36
31	First and foremost, I<AnyEvent is not an event model> itself, it only	37	First and foremost, I<AnyEvent is not an event model> itself, it only
32	interfaces to whatever event model the main program happens to use in a	38	interfaces to whatever event model the main program happens to use in a
33	pragmatic way. For event models and certain classes of immortals alike,	39	pragmatic way. For event models and certain classes of immortals alike,
34	the statement "there can only be one" is a bitter reality, and AnyEvent	40	the statement "there can only be one" is a bitter reality: In general,
35	helps hiding the differences.	41	only one event loop can be active at the same time in a process. AnyEvent
		42	helps hiding the differences between those event loops.
36		43
37	The goal of AnyEvent is to offer module authors the ability to do event	44	The goal of AnyEvent is to offer module authors the ability to do event
38	programming (waiting for I/O or timer events) without subscribing to a	45	programming (waiting for I/O or timer events) without subscribing to a
39	religion, a way of living, and most importantly: without forcing your	46	religion, a way of living, and most importantly: without forcing your
40	module users into the same thing by forcing them to use the same event	47	module users into the same thing by forcing them to use the same event
41	model you use.	48	model you use.
42		49
43	For modules like POE or IO::Async (which is actually doing all I/O	50	For modules like POE or IO::Async (which is a total misnomer as it is
44	I<synchronously>...), using them in your module is like joining a	51	actually doing all I/O I<synchronously>...), using them in your module is
45	cult: After you joined, you are dependent on them and you cannot use	52	like joining a cult: After you joined, you are dependent on them and you
46	anything else, as it is simply incompatible to everything that isn't	53	cannot use anything else, as it is simply incompatible to everything that
47	itself.	54	isn't itself. What's worse, all the potential users of your module are
		55	I<also> forced to use the same event loop you use.
48		56
49	AnyEvent + POE works fine. AnyEvent + Glib works fine. AnyEvent + Tk	57	AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works
50	works fine etc. etc. but none of these work together with the rest: POE	58	fine. AnyEvent + Tk works fine etc. etc. but none of these work together
51	+ IO::Async? no go. Tk + Event? no go. If your module uses one of	59	with the rest: POE + IO::Async? No go. Tk + Event? No go. Again: if
52	those, every user of your module has to use it, too. If your module	60	your module uses one of those, every user of your module has to use it,
53	uses AnyEvent, it works transparently with all event models it supports	61	too. But if your module uses AnyEvent, it works transparently with all
54	(including stuff like POE and IO::Async).	62	event models it supports (including stuff like POE and IO::Async, as long
		63	as those use one of the supported event loops. It is trivial to add new
		64	event loops to AnyEvent, too, so it is future-proof).
55		65
56	In addition of being free of having to use I<the one and only true event	66	In addition to being free of having to use I<the one and only true event
57	model>, AnyEvent also is free of bloat and policy: with POE or similar	67	model>, AnyEvent also is free of bloat and policy: with POE or similar
58	modules, you get an enourmous amount of code and strict rules you have	68	modules, you get an enormous amount of code and strict rules you have to
59	to follow. AnyEvent, on the other hand, is lean and to the point by only	69	follow. AnyEvent, on the other hand, is lean and up to the point, by only
60	offering the functionality that is useful, in as thin as a wrapper as	70	offering the functionality that is necessary, in as thin as a wrapper as
61	technically possible.	71	technically possible.
62		72
		73	Of course, AnyEvent comes with a big (and fully optional!) toolbox
		74	of useful functionality, such as an asynchronous DNS resolver, 100%
		75	non-blocking connects (even with TLS/SSL, IPv6 and on broken platforms
		76	such as Windows) and lots of real-world knowledge and workarounds for
		77	platform bugs and differences.
		78
63	Of course, if you want lots of policy (this is arguably somewhat useful	79	Now, if you I<do want> lots of policy (this can arguably be somewhat
64	in many cases) and you want to force your users to the one and only event	80	useful) and you want to force your users to use the one and only event
65	model your module forces on them, you should I<not> use this module.	81	model, you should I<not> use this module.
66
67		82
68	=head1 DESCRIPTION	83	=head1 DESCRIPTION
69		84
70	L<AnyEvent> provides an identical interface to multiple event loops. This	85	L<AnyEvent> provides an identical interface to multiple event loops. This
71	allows module authors to utilise an event loop without forcing module	86	allows module authors to utilise an event loop without forcing module
72	users to use the same event loop (as only a single event loop can coexist	87	users to use the same event loop (as only a single event loop can coexist
73	peacefully at any one time).	88	peacefully at any one time).
74		89
75	The interface itself is vaguely similar but not identical to the Event	90	The interface itself is vaguely similar, but not identical to the L<Event>
76	module.	91	module.
77		92
78	On the first call of any method, the module tries to detect the currently	93	During the first call of any watcher-creation method, the module tries
79	loaded event loop by probing wether any of the following modules is	94	to detect the currently loaded event loop by probing whether one of the
80	loaded: L<Coro::Event>, L<Event>, L<Glib>, L<Tk>. The first one found is	95	following modules is already loaded: L<EV>,
81	used. If none is found, the module tries to load these modules in the	96	L<Event>, L<Glib>, L<AnyEvent::Impl::Perl>, L<Tk>, L<Event::Lib>, L<Qt>,
82	order given. The first one that could be successfully loaded will be	97	L<POE>. The first one found is used. If none are found, the module tries
83	used. If still none could be found, AnyEvent will fall back to a pure-perl	98	to load these modules (excluding Tk, Event::Lib, Qt and POE as the pure perl
84	event loop, which is also not very efficient.	99	adaptor should always succeed) in the order given. The first one that can
		100	be successfully loaded will be used. If, after this, still none could be
		101	found, AnyEvent will fall back to a pure-perl event loop, which is not
		102	very efficient, but should work everywhere.
85		103
86	Because AnyEvent first checks for modules that are already loaded, loading	104	Because AnyEvent first checks for modules that are already loaded, loading
87	an Event model explicitly before first using AnyEvent will likely make	105	an event model explicitly before first using AnyEvent will likely make
88	that model the default. For example:	106	that model the default. For example:
89		107
90	use Tk;	108	use Tk;
91	use AnyEvent;	109	use AnyEvent;
92		110
93	# .. AnyEvent will likely default to Tk	111	# .. AnyEvent will likely default to Tk
94		112
		113	The I<likely> means that, if any module loads another event model and
		114	starts using it, all bets are off. Maybe you should tell their authors to
		115	use AnyEvent so their modules work together with others seamlessly...
		116
95	The pure-perl implementation of AnyEvent is called	117	The pure-perl implementation of AnyEvent is called
96	C<AnyEvent::Impl::Perl>. Like other event modules you can load it	118	C<AnyEvent::Impl::Perl>. Like other event modules you can load it
97	explicitly.	119	explicitly and enjoy the high availability of that event loop :)
98		120
99	=head1 WATCHERS	121	=head1 WATCHERS
100		122
101	AnyEvent has the central concept of a I<watcher>, which is an object that	123	AnyEvent has the central concept of a I<watcher>, which is an object that
102	stores relevant data for each kind of event you are waiting for, such as	124	stores relevant data for each kind of event you are waiting for, such as
103	the callback to call, the filehandle to watch, etc.	125	the callback to call, the file handle to watch, etc.
104		126
105	These watchers are normal Perl objects with normal Perl lifetime. After	127	These watchers are normal Perl objects with normal Perl lifetime. After
106	creating a watcher it will immediately "watch" for events and invoke	128	creating a watcher it will immediately "watch" for events and invoke the
		129	callback when the event occurs (of course, only when the event model
		130	is in control).
		131
107	the callback. To disable the watcher you have to destroy it (e.g. by	132	To disable the watcher you have to destroy it (e.g. by setting the
108	setting the variable that stores it to C<undef> or otherwise deleting all	133	variable you store it in to C<undef> or otherwise deleting all references
109	references to it).	134	to it).
110		135
111	All watchers are created by calling a method on the C<AnyEvent> class.	136	All watchers are created by calling a method on the C<AnyEvent> class.
112		137
		138	Many watchers either are used with "recursion" (repeating timers for
		139	example), or need to refer to their watcher object in other ways.
		140
		141	An any way to achieve that is this pattern:
		142
		143	my $w; $w = AnyEvent->type (arg => value ..., cb => sub {
		144	# you can use $w here, for example to undef it
		145	undef $w;
		146	});
		147
		148	Note that C<my $w; $w => combination. This is necessary because in Perl,
		149	my variables are only visible after the statement in which they are
		150	declared.
		151
113	=head2 IO WATCHERS	152	=head2 I/O WATCHERS
114		153
115	You can create I/O watcher by calling the C<< AnyEvent->io >> method with	154	You can create an I/O watcher by calling the C<< AnyEvent->io >> method
116	the following mandatory arguments:	155	with the following mandatory key-value pairs as arguments:
117		156
118	C<fh> the Perl I<filehandle> (not filedescriptor) to watch for	157	C<fh> the Perl I<file handle> (I<not> file descriptor) to watch
119	events. C<poll> must be a string that is either C<r> or C<w>, that creates	158	for events. C<poll> must be a string that is either C<r> or C<w>,
120	a watcher waiting for "r"eadable or "w"ritable events. C<cb> the callback	159	which creates a watcher waiting for "r"eadable or "w"ritable events,
121	to invoke everytime the filehandle becomes ready.	160	respectively. C<cb> is the callback to invoke each time the file handle
		161	becomes ready.
122		162
123	Only one io watcher per C<fh> and C<poll> combination is allowed (i.e. on	163	Although the callback might get passed parameters, their value and
124	a socket you can have one r + one w, not any more (limitation comes from	164	presence is undefined and you cannot rely on them. Portable AnyEvent
125	Tk - if you are sure you are not using Tk this limitation is gone).	165	callbacks cannot use arguments passed to I/O watcher callbacks.
126		166
127	Filehandles will be kept alive, so as long as the watcher exists, the	167	The I/O watcher might use the underlying file descriptor or a copy of it.
128	filehandle exists, too.	168	You must not close a file handle as long as any watcher is active on the
		169	underlying file descriptor.
129		170
130	Example:	171	Some event loops issue spurious readyness notifications, so you should
		172	always use non-blocking calls when reading/writing from/to your file
		173	handles.
131		174
132	# wait for readability of STDIN, then read a line and disable the watcher	175	Example: wait for readability of STDIN, then read a line and disable the
		176	watcher.
		177
133	my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {	178	my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
134	chomp (my $input = <STDIN>);	179	chomp (my $input = <STDIN>);
135	warn "read: $input\n";	180	warn "read: $input\n";
136	undef $w;	181	undef $w;
137	});	182	});
…		…
139	=head2 TIME WATCHERS	184	=head2 TIME WATCHERS
140		185
141	You can create a time watcher by calling the C<< AnyEvent->timer >>	186	You can create a time watcher by calling the C<< AnyEvent->timer >>
142	method with the following mandatory arguments:	187	method with the following mandatory arguments:
143		188
144	C<after> after how many seconds (fractions are supported) should the timer	189	C<after> specifies after how many seconds (fractional values are
145	activate. C<cb> the callback to invoke.	190	supported) the callback should be invoked. C<cb> is the callback to invoke
		191	in that case.
146		192
147	The timer callback will be invoked at most once: if you want a repeating	193	Although the callback might get passed parameters, their value and
148	timer you have to create a new watcher (this is a limitation by both Tk	194	presence is undefined and you cannot rely on them. Portable AnyEvent
149	and Glib).	195	callbacks cannot use arguments passed to time watcher callbacks.
150		196
151	Example:	197	The callback will normally be invoked once only. If you specify another
		198	parameter, C<interval>, as a strictly positive number (> 0), then the
		199	callback will be invoked regularly at that interval (in fractional
		200	seconds) after the first invocation. If C<interval> is specified with a
		201	false value, then it is treated as if it were missing.
152		202
		203	The callback will be rescheduled before invoking the callback, but no
		204	attempt is done to avoid timer drift in most backends, so the interval is
		205	only approximate.
		206
153	# fire an event after 7.7 seconds	207	Example: fire an event after 7.7 seconds.
		208
154	my $w = AnyEvent->timer (after => 7.7, cb => sub {	209	my $w = AnyEvent->timer (after => 7.7, cb => sub {
155	warn "timeout\n";	210	warn "timeout\n";
156	});	211	});
157		212
158	# to cancel the timer:	213	# to cancel the timer:
159	undef $w;	214	undef $w;
160		215
161	=head2 CONDITION WATCHERS	216	Example 2: fire an event after 0.5 seconds, then roughly every second.
162		217
163	Condition watchers can be created by calling the C<< AnyEvent->condvar >>	218	my $w = AnyEvent->timer (after => 0.5, interval => 1, cb => sub {
164	method without any arguments.	219	warn "timeout\n";
		220	};
165		221
166	A condition watcher watches for a condition - precisely that the C<<	222	=head3 TIMING ISSUES
167	->broadcast >> method has been called.
168		223
169	Note that condition watchers recurse into the event loop - if you have	224	There are two ways to handle timers: based on real time (relative, "fire
170	two watchers that call C<< ->wait >> in a round-robbin fashion, you	225	in 10 seconds") and based on wallclock time (absolute, "fire at 12
171	lose. Therefore, condition watchers are good to export to your caller, but	226	o'clock").
172	you should avoid making a blocking wait, at least in callbacks, as this
173	usually asks for trouble.
174		227
175	The watcher has only two methods:	228	While most event loops expect timers to specified in a relative way, they
		229	use absolute time internally. This makes a difference when your clock
		230	"jumps", for example, when ntp decides to set your clock backwards from
		231	the wrong date of 2014-01-01 to 2008-01-01, a watcher that is supposed to
		232	fire "after" a second might actually take six years to finally fire.
		233
		234	AnyEvent cannot compensate for this. The only event loop that is conscious
		235	about these issues is L<EV>, which offers both relative (ev_timer, based
		236	on true relative time) and absolute (ev_periodic, based on wallclock time)
		237	timers.
		238
		239	AnyEvent always prefers relative timers, if available, matching the
		240	AnyEvent API.
		241
		242	AnyEvent has two additional methods that return the "current time":
176		243
177	=over 4	244	=over 4
178		245
179	=item $cv->wait	246	=item AnyEvent->time
180		247
181	Wait (blocking if necessary) until the C<< ->broadcast >> method has been	248	This returns the "current wallclock time" as a fractional number of
182	called on c<$cv>, while servicing other watchers normally.	249	seconds since the Epoch (the same thing as C<time> or C<Time::HiRes::time>
		250	return, and the result is guaranteed to be compatible with those).
183		251
184	Not all event models support a blocking wait - some die in that case, so	252	It progresses independently of any event loop processing, i.e. each call
185	if you are using this from a module, never require a blocking wait, but	253	will check the system clock, which usually gets updated frequently.
186	let the caller decide wether the call will block or not (for example,
187	by coupling condition variables with some kind of request results and
188	supporting callbacks so the caller knows that getting the result will not
189	block, while still suppporting blockign waits if the caller so desires).
190		254
191	You can only wait once on a condition - additional calls will return	255	=item AnyEvent->now
192	immediately.
193		256
194	=item $cv->broadcast	257	This also returns the "current wallclock time", but unlike C<time>, above,
		258	this value might change only once per event loop iteration, depending on
		259	the event loop (most return the same time as C<time>, above). This is the
		260	time that AnyEvent's timers get scheduled against.
195		261
196	Flag the condition as ready - a running C<< ->wait >> and all further	262	I<In almost all cases (in all cases if you don't care), this is the
197	calls to C<wait> will return after this method has been called. If nobody	263	function to call when you want to know the current time.>
198	is waiting the broadcast will be remembered..
199		264
200	Example:	265	This function is also often faster then C<< AnyEvent->time >>, and
		266	thus the preferred method if you want some timestamp (for example,
		267	L<AnyEvent::Handle> uses this to update it's activity timeouts).
		268
		269	The rest of this section is only of relevance if you try to be very exact
		270	with your timing, you can skip it without bad conscience.
		271
		272	For a practical example of when these times differ, consider L<Event::Lib>
		273	and L<EV> and the following set-up:
		274
		275	The event loop is running and has just invoked one of your callback at
		276	time=500 (assume no other callbacks delay processing). In your callback,
		277	you wait a second by executing C<sleep 1> (blocking the process for a
		278	second) and then (at time=501) you create a relative timer that fires
		279	after three seconds.
		280
		281	With L<Event::Lib>, C<< AnyEvent->time >> and C<< AnyEvent->now >> will
		282	both return C<501>, because that is the current time, and the timer will
		283	be scheduled to fire at time=504 (C<501> + C<3>).
		284
		285	With L<EV>, C<< AnyEvent->time >> returns C<501> (as that is the current
		286	time), but C<< AnyEvent->now >> returns C<500>, as that is the time the
		287	last event processing phase started. With L<EV>, your timer gets scheduled
		288	to run at time=503 (C<500> + C<3>).
		289
		290	In one sense, L<Event::Lib> is more exact, as it uses the current time
		291	regardless of any delays introduced by event processing. However, most
		292	callbacks do not expect large delays in processing, so this causes a
		293	higher drift (and a lot more system calls to get the current time).
		294
		295	In another sense, L<EV> is more exact, as your timer will be scheduled at
		296	the same time, regardless of how long event processing actually took.
		297
		298	In either case, if you care (and in most cases, you don't), then you
		299	can get whatever behaviour you want with any event loop, by taking the
		300	difference between C<< AnyEvent->time >> and C<< AnyEvent->now >> into
		301	account.
		302
		303	=back
		304
		305	=head2 SIGNAL WATCHERS
		306
		307	You can watch for signals using a signal watcher, C<signal> is the signal
		308	I<name> without any C<SIG> prefix, C<cb> is the Perl callback to
		309	be invoked whenever a signal occurs.
		310
		311	Although the callback might get passed parameters, their value and
		312	presence is undefined and you cannot rely on them. Portable AnyEvent
		313	callbacks cannot use arguments passed to signal watcher callbacks.
		314
		315	Multiple signal occurrences can be clumped together into one callback
		316	invocation, and callback invocation will be synchronous. Synchronous means
		317	that it might take a while until the signal gets handled by the process,
		318	but it is guaranteed not to interrupt any other callbacks.
		319
		320	The main advantage of using these watchers is that you can share a signal
		321	between multiple watchers.
		322
		323	This watcher might use C<%SIG>, so programs overwriting those signals
		324	directly will likely not work correctly.
		325
		326	Example: exit on SIGINT
		327
		328	my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 });
		329
		330	=head2 CHILD PROCESS WATCHERS
		331
		332	You can also watch on a child process exit and catch its exit status.
		333
		334	The child process is specified by the C<pid> argument (if set to C<0>, it
		335	watches for any child process exit). The watcher will trigger as often
		336	as status change for the child are received. This works by installing a
		337	signal handler for C<SIGCHLD>. The callback will be called with the pid
		338	and exit status (as returned by waitpid), so unlike other watcher types,
		339	you I<can> rely on child watcher callback arguments.
		340
		341	There is a slight catch to child watchers, however: you usually start them
		342	I<after> the child process was created, and this means the process could
		343	have exited already (and no SIGCHLD will be sent anymore).
		344
		345	Not all event models handle this correctly (POE doesn't), but even for
		346	event models that I<do> handle this correctly, they usually need to be
		347	loaded before the process exits (i.e. before you fork in the first place).
		348
		349	This means you cannot create a child watcher as the very first thing in an
		350	AnyEvent program, you I<have> to create at least one watcher before you
		351	C<fork> the child (alternatively, you can call C<AnyEvent::detect>).
		352
		353	Example: fork a process and wait for it
		354
		355	my $done = AnyEvent->condvar;
		356
		357	my $pid = fork or exit 5;
		358
		359	my $w = AnyEvent->child (
		360	pid => $pid,
		361	cb => sub {
		362	my ($pid, $status) = @_;
		363	warn "pid $pid exited with status $status";
		364	$done->send;
		365	},
		366	);
		367
		368	# do something else, then wait for process exit
		369	$done->recv;
		370
		371	=head2 CONDITION VARIABLES
		372
		373	If you are familiar with some event loops you will know that all of them
		374	require you to run some blocking "loop", "run" or similar function that
		375	will actively watch for new events and call your callbacks.
		376
		377	AnyEvent is different, it expects somebody else to run the event loop and
		378	will only block when necessary (usually when told by the user).
		379
		380	The instrument to do that is called a "condition variable", so called
		381	because they represent a condition that must become true.
		382
		383	Condition variables can be created by calling the C<< AnyEvent->condvar
		384	>> method, usually without arguments. The only argument pair allowed is
		385	C<cb>, which specifies a callback to be called when the condition variable
		386	becomes true.
		387
		388	After creation, the condition variable is "false" until it becomes "true"
		389	by calling the C<send> method (or calling the condition variable as if it
		390	were a callback, read about the caveats in the description for the C<<
		391	->send >> method).
		392
		393	Condition variables are similar to callbacks, except that you can
		394	optionally wait for them. They can also be called merge points - points
		395	in time where multiple outstanding events have been processed. And yet
		396	another way to call them is transactions - each condition variable can be
		397	used to represent a transaction, which finishes at some point and delivers
		398	a result.
		399
		400	Condition variables are very useful to signal that something has finished,
		401	for example, if you write a module that does asynchronous http requests,
		402	then a condition variable would be the ideal candidate to signal the
		403	availability of results. The user can either act when the callback is
		404	called or can synchronously C<< ->recv >> for the results.
		405
		406	You can also use them to simulate traditional event loops - for example,
		407	you can block your main program until an event occurs - for example, you
		408	could C<< ->recv >> in your main program until the user clicks the Quit
		409	button of your app, which would C<< ->send >> the "quit" event.
		410
		411	Note that condition variables recurse into the event loop - if you have
		412	two pieces of code that call C<< ->recv >> in a round-robin fashion, you
		413	lose. Therefore, condition variables are good to export to your caller, but
		414	you should avoid making a blocking wait yourself, at least in callbacks,
		415	as this asks for trouble.
		416
		417	Condition variables are represented by hash refs in perl, and the keys
		418	used by AnyEvent itself are all named C<_ae_XXX> to make subclassing
		419	easy (it is often useful to build your own transaction class on top of
		420	AnyEvent). To subclass, use C<AnyEvent::CondVar> as base class and call
		421	it's C<new> method in your own C<new> method.
		422
		423	There are two "sides" to a condition variable - the "producer side" which
		424	eventually calls C<< -> send >>, and the "consumer side", which waits
		425	for the send to occur.
		426
		427	Example: wait for a timer.
201		428
202	# wait till the result is ready	429	# wait till the result is ready
203	my $result_ready = AnyEvent->condvar;	430	my $result_ready = AnyEvent->condvar;
204		431
205	# do something such as adding a timer	432	# do something such as adding a timer
206	# or socket watcher the calls $result_ready->broadcast	433	# or socket watcher the calls $result_ready->send
207	# when the "result" is ready.	434	# when the "result" is ready.
		435	# in this case, we simply use a timer:
		436	my $w = AnyEvent->timer (
		437	after => 1,
		438	cb => sub { $result_ready->send },
		439	);
208		440
		441	# this "blocks" (while handling events) till the callback
		442	# calls send
209	$result_ready->wait;	443	$result_ready->recv;
		444
		445	Example: wait for a timer, but take advantage of the fact that
		446	condition variables are also code references.
		447
		448	my $done = AnyEvent->condvar;
		449	my $delay = AnyEvent->timer (after => 5, cb => $done);
		450	$done->recv;
		451
		452	=head3 METHODS FOR PRODUCERS
		453
		454	These methods should only be used by the producing side, i.e. the
		455	code/module that eventually sends the signal. Note that it is also
		456	the producer side which creates the condvar in most cases, but it isn't
		457	uncommon for the consumer to create it as well.
		458
		459	=over 4
		460
		461	=item $cv->send (...)
		462
		463	Flag the condition as ready - a running C<< ->recv >> and all further
		464	calls to C<recv> will (eventually) return after this method has been
		465	called. If nobody is waiting the send will be remembered.
		466
		467	If a callback has been set on the condition variable, it is called
		468	immediately from within send.
		469
		470	Any arguments passed to the C<send> call will be returned by all
		471	future C<< ->recv >> calls.
		472
		473	Condition variables are overloaded so one can call them directly
		474	(as a code reference). Calling them directly is the same as calling
		475	C<send>. Note, however, that many C-based event loops do not handle
		476	overloading, so as tempting as it may be, passing a condition variable
		477	instead of a callback does not work. Both the pure perl and EV loops
		478	support overloading, however, as well as all functions that use perl to
		479	invoke a callback (as in L<AnyEvent::Socket> and L<AnyEvent::DNS> for
		480	example).
		481
		482	=item $cv->croak ($error)
		483
		484	Similar to send, but causes all call's to C<< ->recv >> to invoke
		485	C<Carp::croak> with the given error message/object/scalar.
		486
		487	This can be used to signal any errors to the condition variable
		488	user/consumer.
		489
		490	=item $cv->begin ([group callback])
		491
		492	=item $cv->end
		493
		494	These two methods are EXPERIMENTAL and MIGHT CHANGE.
		495
		496	These two methods can be used to combine many transactions/events into
		497	one. For example, a function that pings many hosts in parallel might want
		498	to use a condition variable for the whole process.
		499
		500	Every call to C<< ->begin >> will increment a counter, and every call to
		501	C<< ->end >> will decrement it. If the counter reaches C<0> in C<< ->end
		502	>>, the (last) callback passed to C<begin> will be executed. That callback
		503	is I<supposed> to call C<< ->send >>, but that is not required. If no
		504	callback was set, C<send> will be called without any arguments.
		505
		506	Let's clarify this with the ping example:
		507
		508	my $cv = AnyEvent->condvar;
		509
		510	my %result;
		511	$cv->begin (sub { $cv->send (\%result) });
		512
		513	for my $host (@list_of_hosts) {
		514	$cv->begin;
		515	ping_host_then_call_callback $host, sub {
		516	$result{$host} = ...;
		517	$cv->end;
		518	};
		519	}
		520
		521	$cv->end;
		522
		523	This code fragment supposedly pings a number of hosts and calls
		524	C<send> after results for all then have have been gathered - in any
		525	order. To achieve this, the code issues a call to C<begin> when it starts
		526	each ping request and calls C<end> when it has received some result for
		527	it. Since C<begin> and C<end> only maintain a counter, the order in which
		528	results arrive is not relevant.
		529
		530	There is an additional bracketing call to C<begin> and C<end> outside the
		531	loop, which serves two important purposes: first, it sets the callback
		532	to be called once the counter reaches C<0>, and second, it ensures that
		533	C<send> is called even when C<no> hosts are being pinged (the loop
		534	doesn't execute once).
		535
		536	This is the general pattern when you "fan out" into multiple subrequests:
		537	use an outer C<begin>/C<end> pair to set the callback and ensure C<end>
		538	is called at least once, and then, for each subrequest you start, call
		539	C<begin> and for each subrequest you finish, call C<end>.
210		540
211	=back	541	=back
212		542
213	=head2 SIGNAL WATCHERS	543	=head3 METHODS FOR CONSUMERS
214		544
215	You can listen for signals using a signal watcher, C<signal> is the signal	545	These methods should only be used by the consuming side, i.e. the
216	I<name> without any C<SIG> prefix. Multiple signals events can be clumped	546	code awaits the condition.
217	together into one callback invocation, and callback invocation might or
218	might not be asynchronous.
219		547
220	These watchers might use C<%SIG>, so programs overwriting those signals	548	=over 4
221	directly will likely not work correctly.
222		549
223	Example: exit on SIGINT	550	=item $cv->recv
224		551
225	my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 });	552	Wait (blocking if necessary) until the C<< ->send >> or C<< ->croak
		553	>> methods have been called on c<$cv>, while servicing other watchers
		554	normally.
226		555
227	=head2 CHILD PROCESS WATCHERS	556	You can only wait once on a condition - additional calls are valid but
		557	will return immediately.
228		558
229	You can also listen for the status of a child process specified by the	559	If an error condition has been set by calling C<< ->croak >>, then this
230	C<pid> argument (or any child if the pid argument is 0). The watcher will	560	function will call C<croak>.
231	trigger as often as status change for the child are received. This works
232	by installing a signal handler for C<SIGCHLD>. The callback will be called with
233	the pid and exit status (as returned by waitpid).
234		561
235	Example: wait for pid 1333	562	In list context, all parameters passed to C<send> will be returned,
		563	in scalar context only the first one will be returned.
236		564
237	my $w = AnyEvent->child (pid => 1333, cb => sub { warn "exit status $?" });	565	Not all event models support a blocking wait - some die in that case
		566	(programs might want to do that to stay interactive), so I<if you are
		567	using this from a module, never require a blocking wait>, but let the
		568	caller decide whether the call will block or not (for example, by coupling
		569	condition variables with some kind of request results and supporting
		570	callbacks so the caller knows that getting the result will not block,
		571	while still supporting blocking waits if the caller so desires).
238		572
239	=head1 GLOBALS	573	Another reason I<never> to C<< ->recv >> in a module is that you cannot
		574	sensibly have two C<< ->recv >>'s in parallel, as that would require
		575	multiple interpreters or coroutines/threads, none of which C<AnyEvent>
		576	can supply.
		577
		578	The L<Coro> module, however, I<can> and I<does> supply coroutines and, in
		579	fact, L<Coro::AnyEvent> replaces AnyEvent's condvars by coroutine-safe
		580	versions and also integrates coroutines into AnyEvent, making blocking
		581	C<< ->recv >> calls perfectly safe as long as they are done from another
		582	coroutine (one that doesn't run the event loop).
		583
		584	You can ensure that C<< -recv >> never blocks by setting a callback and
		585	only calling C<< ->recv >> from within that callback (or at a later
		586	time). This will work even when the event loop does not support blocking
		587	waits otherwise.
		588
		589	=item $bool = $cv->ready
		590
		591	Returns true when the condition is "true", i.e. whether C<send> or
		592	C<croak> have been called.
		593
		594	=item $cb = $cv->cb ([new callback])
		595
		596	This is a mutator function that returns the callback set and optionally
		597	replaces it before doing so.
		598
		599	The callback will be called when the condition becomes "true", i.e. when
		600	C<send> or C<croak> are called, with the only argument being the condition
		601	variable itself. Calling C<recv> inside the callback or at any later time
		602	is guaranteed not to block.
		603
		604	=back
		605
		606	=head1 GLOBAL VARIABLES AND FUNCTIONS
240		607
241	=over 4	608	=over 4
242		609
243	=item $AnyEvent::MODEL	610	=item $AnyEvent::MODEL
244		611
…		…
248	C<AnyEvent::Impl:xxx> modules, but can be any other class in the case	615	C<AnyEvent::Impl:xxx> modules, but can be any other class in the case
249	AnyEvent has been extended at runtime (e.g. in I<rxvt-unicode>).	616	AnyEvent has been extended at runtime (e.g. in I<rxvt-unicode>).
250		617
251	The known classes so far are:	618	The known classes so far are:
252		619
253	AnyEvent::Impl::CoroEV based on Coro::EV, best choice.
254	AnyEvent::Impl::EV based on EV (an interface to libev, also best choice).	620	AnyEvent::Impl::EV based on EV (an interface to libev, best choice).
255	AnyEvent::Impl::CoroEvent based on Coro::Event, second best choice.
256	AnyEvent::Impl::Event based on Event, also second best choice :)	621	AnyEvent::Impl::Event based on Event, second best choice.
		622	AnyEvent::Impl::Perl pure-perl implementation, fast and portable.
257	AnyEvent::Impl::Glib based on Glib, second-best choice.	623	AnyEvent::Impl::Glib based on Glib, third-best choice.
258	AnyEvent::Impl::Tk based on Tk, very bad choice.	624	AnyEvent::Impl::Tk based on Tk, very bad choice.
259	AnyEvent::Impl::Perl pure-perl implementation, inefficient.	625	AnyEvent::Impl::Qt based on Qt, cannot be autoprobed (see its docs).
		626	AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse.
		627	AnyEvent::Impl::POE based on POE, not generic enough for full support.
		628
		629	There is no support for WxWidgets, as WxWidgets has no support for
		630	watching file handles. However, you can use WxWidgets through the
		631	POE Adaptor, as POE has a Wx backend that simply polls 20 times per
		632	second, which was considered to be too horrible to even consider for
		633	AnyEvent. Likewise, other POE backends can be used by AnyEvent by using
		634	it's adaptor.
		635
		636	AnyEvent knows about L<Prima> and L<Wx> and will try to use L<POE> when
		637	autodetecting them.
260		638
261	=item AnyEvent::detect	639	=item AnyEvent::detect
262		640
263	Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model if	641	Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model
264	necessary. You should only call this function right before you would have	642	if necessary. You should only call this function right before you would
265	created an AnyEvent watcher anyway, that is, very late at runtime.	643	have created an AnyEvent watcher anyway, that is, as late as possible at
		644	runtime.
		645
		646	=item $guard = AnyEvent::post_detect { BLOCK }
		647
		648	Arranges for the code block to be executed as soon as the event model is
		649	autodetected (or immediately if this has already happened).
		650
		651	If called in scalar or list context, then it creates and returns an object
		652	that automatically removes the callback again when it is destroyed. See
		653	L<Coro::BDB> for a case where this is useful.
		654
		655	=item @AnyEvent::post_detect
		656
		657	If there are any code references in this array (you can C<push> to it
		658	before or after loading AnyEvent), then they will called directly after
		659	the event loop has been chosen.
		660
		661	You should check C<$AnyEvent::MODEL> before adding to this array, though:
		662	if it contains a true value then the event loop has already been detected,
		663	and the array will be ignored.
		664
		665	Best use C<AnyEvent::post_detect { BLOCK }> instead.
266		666
267	=back	667	=back
268		668
269	=head1 WHAT TO DO IN A MODULE	669	=head1 WHAT TO DO IN A MODULE
270		670
271	As a module author, you should "use AnyEvent" and call AnyEvent methods	671	As a module author, you should C<use AnyEvent> and call AnyEvent methods
272	freely, but you should not load a specific event module or rely on it.	672	freely, but you should not load a specific event module or rely on it.
273		673
274	Be careful when you create watchers in the module body - Anyevent will	674	Be careful when you create watchers in the module body - AnyEvent will
275	decide which event module to use as soon as the first method is called, so	675	decide which event module to use as soon as the first method is called, so
276	by calling AnyEvent in your module body you force the user of your module	676	by calling AnyEvent in your module body you force the user of your module
277	to load the event module first.	677	to load the event module first.
278		678
		679	Never call C<< ->recv >> on a condition variable unless you I<know> that
		680	the C<< ->send >> method has been called on it already. This is
		681	because it will stall the whole program, and the whole point of using
		682	events is to stay interactive.
		683
		684	It is fine, however, to call C<< ->recv >> when the user of your module
		685	requests it (i.e. if you create a http request object ad have a method
		686	called C<results> that returns the results, it should call C<< ->recv >>
		687	freely, as the user of your module knows what she is doing. always).
		688
279	=head1 WHAT TO DO IN THE MAIN PROGRAM	689	=head1 WHAT TO DO IN THE MAIN PROGRAM
280		690
281	There will always be a single main program - the only place that should	691	There will always be a single main program - the only place that should
282	dictate which event model to use.	692	dictate which event model to use.
283		693
284	If it doesn't care, it can just "use AnyEvent" and use it itself, or not	694	If it doesn't care, it can just "use AnyEvent" and use it itself, or not
285	do anything special and let AnyEvent decide which implementation to chose.	695	do anything special (it does not need to be event-based) and let AnyEvent
		696	decide which implementation to chose if some module relies on it.
286		697
287	If the main program relies on a specific event model (for example, in Gtk2	698	If the main program relies on a specific event model - for example, in
288	programs you have to rely on either Glib or Glib::Event), you should load	699	Gtk2 programs you have to rely on the Glib module - you should load the
289	it before loading AnyEvent or any module that uses it, generally, as early	700	event module before loading AnyEvent or any module that uses it: generally
290	as possible. The reason is that modules might create watchers when they	701	speaking, you should load it as early as possible. The reason is that
291	are loaded, and AnyEvent will decide on the event model to use as soon as	702	modules might create watchers when they are loaded, and AnyEvent will
292	it creates watchers, and it might chose the wrong one unless you load the	703	decide on the event model to use as soon as it creates watchers, and it
293	correct one yourself.	704	might chose the wrong one unless you load the correct one yourself.
294		705
295	You can chose to use a rather inefficient pure-perl implementation by	706	You can chose to use a pure-perl implementation by loading the
296	loading the C<AnyEvent::Impl::Perl> module, but letting AnyEvent chose is	707	C<AnyEvent::Impl::Perl> module, which gives you similar behaviour
297	generally better.	708	everywhere, but letting AnyEvent chose the model is generally better.
		709
		710	=head2 MAINLOOP EMULATION
		711
		712	Sometimes (often for short test scripts, or even standalone programs who
		713	only want to use AnyEvent), you do not want to run a specific event loop.
		714
		715	In that case, you can use a condition variable like this:
		716
		717	AnyEvent->condvar->recv;
		718
		719	This has the effect of entering the event loop and looping forever.
		720
		721	Note that usually your program has some exit condition, in which case
		722	it is better to use the "traditional" approach of storing a condition
		723	variable somewhere, waiting for it, and sending it when the program should
		724	exit cleanly.
		725
		726
		727	=head1 OTHER MODULES
		728
		729	The following is a non-exhaustive list of additional modules that use
		730	AnyEvent and can therefore be mixed easily with other AnyEvent modules
		731	in the same program. Some of the modules come with AnyEvent, some are
		732	available via CPAN.
		733
		734	=over 4
		735
		736	=item L<AnyEvent::Util>
		737
		738	Contains various utility functions that replace often-used but blocking
		739	functions such as C<inet_aton> by event-/callback-based versions.
		740
		741	=item L<AnyEvent::Socket>
		742
		743	Provides various utility functions for (internet protocol) sockets,
		744	addresses and name resolution. Also functions to create non-blocking tcp
		745	connections or tcp servers, with IPv6 and SRV record support and more.
		746
		747	=item L<AnyEvent::Handle>
		748
		749	Provide read and write buffers, manages watchers for reads and writes,
		750	supports raw and formatted I/O, I/O queued and fully transparent and
		751	non-blocking SSL/TLS.
		752
		753	=item L<AnyEvent::DNS>
		754
		755	Provides rich asynchronous DNS resolver capabilities.
		756
		757	=item L<AnyEvent::HTTP>
		758
		759	A simple-to-use HTTP library that is capable of making a lot of concurrent
		760	HTTP requests.
		761
		762	=item L<AnyEvent::HTTPD>
		763
		764	Provides a simple web application server framework.
		765
		766	=item L<AnyEvent::FastPing>
		767
		768	The fastest ping in the west.
		769
		770	=item L<AnyEvent::DBI>
		771
		772	Executes L<DBI> requests asynchronously in a proxy process.
		773
		774	=item L<AnyEvent::AIO>
		775
		776	Truly asynchronous I/O, should be in the toolbox of every event
		777	programmer. AnyEvent::AIO transparently fuses L<IO::AIO> and AnyEvent
		778	together.
		779
		780	=item L<AnyEvent::BDB>
		781
		782	Truly asynchronous Berkeley DB access. AnyEvent::BDB transparently fuses
		783	L<BDB> and AnyEvent together.
		784
		785	=item L<AnyEvent::GPSD>
		786
		787	A non-blocking interface to gpsd, a daemon delivering GPS information.
		788
		789	=item L<AnyEvent::IGS>
		790
		791	A non-blocking interface to the Internet Go Server protocol (used by
		792	L<App::IGS>).
		793
		794	=item L<Net::IRC3>
		795
		796	AnyEvent based IRC client module family.
		797
		798	=item L<Net::XMPP2>
		799
		800	AnyEvent based XMPP (Jabber protocol) module family.
		801
		802	=item L<Net::FCP>
		803
		804	AnyEvent-based implementation of the Freenet Client Protocol, birthplace
		805	of AnyEvent.
		806
		807	=item L<Event::ExecFlow>
		808
		809	High level API for event-based execution flow control.
		810
		811	=item L<Coro>
		812
		813	Has special support for AnyEvent via L<Coro::AnyEvent>.
		814
		815	=item L<IO::Lambda>
		816
		817	The lambda approach to I/O - don't ask, look there. Can use AnyEvent.
		818
		819	=back
298		820
299	=cut	821	=cut
300		822
301	package AnyEvent;	823	package AnyEvent;
302		824
303	no warnings;	825	no warnings;
304	use strict;	826	use strict;
305		827
306	use Carp;	828	use Carp;
307		829
308	our $VERSION = '3.0';	830	our $VERSION = 4.2;
309	our $MODEL;	831	our $MODEL;
310		832
311	our $AUTOLOAD;	833	our $AUTOLOAD;
312	our @ISA;	834	our @ISA;
313		835
		836	our @REGISTRY;
		837
		838	our $WIN32;
		839
		840	BEGIN {
		841	my $win32 = ! ! ($^O =~ /mswin32/i);
		842	eval "sub WIN32(){ $win32 }";
		843	}
		844
314	our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1;	845	our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1;
315		846
316	our @REGISTRY;	847	our %PROTOCOL; # (ipv4|ipv6) => (1|2), higher numbers are preferred
		848
		849	{
		850	my $idx;
		851	$PROTOCOL{$_} = ++$idx
		852	for reverse split /\s*,\s*/,
		853	$ENV{PERL_ANYEVENT_PROTOCOLS} || "ipv4,ipv6";
		854	}
317		855
318	my @models = (	856	my @models = (
319	[Coro::EV:: => AnyEvent::Impl::CoroEV::],
320	[EV:: => AnyEvent::Impl::EV::],	857	[EV:: => AnyEvent::Impl::EV::],
321	[Coro::Event:: => AnyEvent::Impl::CoroEvent::],
322	[Event:: => AnyEvent::Impl::Event::],	858	[Event:: => AnyEvent::Impl::Event::],
323	[Glib:: => AnyEvent::Impl::Glib::],
324	[Tk:: => AnyEvent::Impl::Tk::],
325	[AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl::],	859	[AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl::],
		860	# everything below here will not be autoprobed
		861	# as the pureperl backend should work everywhere
		862	# and is usually faster
		863	[Tk:: => AnyEvent::Impl::Tk::], # crashes with many handles
		864	[Glib:: => AnyEvent::Impl::Glib::], # becomes extremely slow with many watchers
		865	[Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy
		866	[Qt:: => AnyEvent::Impl::Qt::], # requires special main program
		867	[POE::Kernel:: => AnyEvent::Impl::POE::], # lasciate ogni speranza
		868	[Wx:: => AnyEvent::Impl::POE::],
		869	[Prima:: => AnyEvent::Impl::POE::],
326	);	870	);
327		871
328	our %method = map +($_ => 1), qw(io timer condvar broadcast wait signal one_event DESTROY);	872	our %method = map +($_ => 1), qw(io timer time now signal child condvar one_event DESTROY);
		873
		874	our @post_detect;
		875
		876	sub post_detect(&) {
		877	my ($cb) = @_;
		878
		879	if ($MODEL) {
		880	$cb->();
		881
		882	1
		883	} else {
		884	push @post_detect, $cb;
		885
		886	defined wantarray
		887	? bless \$cb, "AnyEvent::Util::PostDetect"
		888	: ()
		889	}
		890	}
		891
		892	sub AnyEvent::Util::PostDetect::DESTROY {
		893	@post_detect = grep $_ != ${$_[0]}, @post_detect;
		894	}
329		895
330	sub detect() {	896	sub detect() {
331	unless ($MODEL) {	897	unless ($MODEL) {
332	no strict 'refs';	898	no strict 'refs';
		899	local $SIG{__DIE__};
		900
		901	if ($ENV{PERL_ANYEVENT_MODEL} =~ /^([a-zA-Z]+)$/) {
		902	my $model = "AnyEvent::Impl::$1";
		903	if (eval "require $model") {
		904	$MODEL = $model;
		905	warn "AnyEvent: loaded model '$model' (forced by \$PERL_ANYEVENT_MODEL), using it.\n" if $verbose > 1;
		906	} else {
		907	warn "AnyEvent: unable to load model '$model' (from \$PERL_ANYEVENT_MODEL):\n$@" if $verbose;
		908	}
		909	}
333		910
334	# check for already loaded models	911	# check for already loaded models
		912	unless ($MODEL) {
335	for (@REGISTRY, @models) {	913	for (@REGISTRY, @models) {
336	my ($package, $model) = @$_;	914	my ($package, $model) = @$_;
337	if (${"$package\::VERSION"} > 0) {	915	if (${"$package\::VERSION"} > 0) {
338	if (eval "require $model") {	916	if (eval "require $model") {
339	$MODEL = $model;	917	$MODEL = $model;
340	warn "AnyEvent: found model '$model', using it.\n" if $verbose > 1;	918	warn "AnyEvent: autodetected model '$model', using it.\n" if $verbose > 1;
341	last;	919	last;
		920	}
342	}	921	}
343	}	922	}
344	}
345		923
346	unless ($MODEL) {	924	unless ($MODEL) {
347	# try to load a model	925	# try to load a model
348		926
349	for (@REGISTRY, @models) {	927	for (@REGISTRY, @models) {
350	my ($package, $model) = @$_;	928	my ($package, $model) = @$_;
351	if (eval "require $package"	929	if (eval "require $package"
352	and ${"$package\::VERSION"} > 0	930	and ${"$package\::VERSION"} > 0
353	and eval "require $model") {	931	and eval "require $model") {
354	$MODEL = $model;	932	$MODEL = $model;
355	warn "AnyEvent: autoprobed and loaded model '$model', using it.\n" if $verbose > 1;	933	warn "AnyEvent: autoprobed model '$model', using it.\n" if $verbose > 1;
356	last;	934	last;
		935	}
357	}	936	}
		937
		938	$MODEL
		939	or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: EV, Event or Glib.";
358	}	940	}
359
360	$MODEL
361	or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: EV (or Coro+EV), Event (or Coro+Event), Glib or Tk.";
362	}	941	}
363		942
364	unshift @ISA, $MODEL;	943	unshift @ISA, $MODEL;
365	push @{"$MODEL\::ISA"}, "AnyEvent::Base";	944	push @{"$MODEL\::ISA"}, "AnyEvent::Base";
		945
		946	(shift @post_detect)->() while @post_detect;
366	}	947	}
367		948
368	$MODEL	949	$MODEL
369	}	950	}
370		951
…		…
380	$class->$func (@_);	961	$class->$func (@_);
381	}	962	}
382		963
383	package AnyEvent::Base;	964	package AnyEvent::Base;
384		965
		966	# default implementation for now and time
		967
		968	use Time::HiRes ();
		969
		970	sub time { Time::HiRes::time }
		971	sub now { Time::HiRes::time }
		972
385	# default implementation for ->condvar, ->wait, ->broadcast	973	# default implementation for ->condvar
386		974
387	sub condvar {	975	sub condvar {
388	bless \my $flag, "AnyEvent::Base::CondVar"	976	bless { @_ == 3 ? (_ae_cb => $_[2]) : () }, AnyEvent::CondVar::
389	}
390
391	sub AnyEvent::Base::CondVar::broadcast {
392	${$_[0]}++;
393	}
394
395	sub AnyEvent::Base::CondVar::wait {
396	AnyEvent->one_event while !${$_[0]};
397	}	977	}
398		978
399	# default implementation for ->signal	979	# default implementation for ->signal
400		980
401	our %SIG_CB;	981	our %SIG_CB;
…		…
417	sub AnyEvent::Base::Signal::DESTROY {	997	sub AnyEvent::Base::Signal::DESTROY {
418	my ($signal, $cb) = @{$_[0]};	998	my ($signal, $cb) = @{$_[0]};
419		999
420	delete $SIG_CB{$signal}{$cb};	1000	delete $SIG_CB{$signal}{$cb};
421		1001
422	$SIG{$signal} = 'DEFAULT' unless keys %{ $SIG_CB{$signal} };	1002	delete $SIG{$signal} unless keys %{ $SIG_CB{$signal} };
423	}	1003	}
424		1004
425	# default implementation for ->child	1005	# default implementation for ->child
426		1006
427	our %PID_CB;	1007	our %PID_CB;
…		…
454	or Carp::croak "required option 'pid' is missing";	1034	or Carp::croak "required option 'pid' is missing";
455		1035
456	$PID_CB{$pid}{$arg{cb}} = $arg{cb};	1036	$PID_CB{$pid}{$arg{cb}} = $arg{cb};
457		1037
458	unless ($WNOHANG) {	1038	unless ($WNOHANG) {
459	$WNOHANG = eval { require POSIX; &POSIX::WNOHANG } || 1;	1039	$WNOHANG = eval { local $SIG{__DIE__}; require POSIX; &POSIX::WNOHANG } || 1;
460	}	1040	}
461		1041
462	unless ($CHLD_W) {	1042	unless ($CHLD_W) {
463	$CHLD_W = AnyEvent->signal (signal => 'CHLD', cb => \&_sigchld);	1043	$CHLD_W = AnyEvent->signal (signal => 'CHLD', cb => \&_sigchld);
464	# child could be a zombie already, so make at least one round	1044	# child could be a zombie already, so make at least one round
…		…
475	delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} };	1055	delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} };
476		1056
477	undef $CHLD_W unless keys %PID_CB;	1057	undef $CHLD_W unless keys %PID_CB;
478	}	1058	}
479		1059
		1060	package AnyEvent::CondVar;
		1061
		1062	our @ISA = AnyEvent::CondVar::Base::;
		1063
		1064	package AnyEvent::CondVar::Base;
		1065
		1066	use overload
		1067	'&{}' => sub { my $self = shift; sub { $self->send (@_) } },
		1068	fallback => 1;
		1069
		1070	sub _send {
		1071	# nop
		1072	}
		1073
		1074	sub send {
		1075	my $cv = shift;
		1076	$cv->{_ae_sent} = [@_];
		1077	(delete $cv->{_ae_cb})->($cv) if $cv->{_ae_cb};
		1078	$cv->_send;
		1079	}
		1080
		1081	sub croak {
		1082	$_[0]{_ae_croak} = $_[1];
		1083	$_[0]->send;
		1084	}
		1085
		1086	sub ready {
		1087	$_[0]{_ae_sent}
		1088	}
		1089
		1090	sub _wait {
		1091	AnyEvent->one_event while !$_[0]{_ae_sent};
		1092	}
		1093
		1094	sub recv {
		1095	$_[0]->_wait;
		1096
		1097	Carp::croak $_[0]{_ae_croak} if $_[0]{_ae_croak};
		1098	wantarray ? @{ $_[0]{_ae_sent} } : $_[0]{_ae_sent}[0]
		1099	}
		1100
		1101	sub cb {
		1102	$_[0]{_ae_cb} = $_[1] if @_ > 1;
		1103	$_[0]{_ae_cb}
		1104	}
		1105
		1106	sub begin {
		1107	++$_[0]{_ae_counter};
		1108	$_[0]{_ae_end_cb} = $_[1] if @_ > 1;
		1109	}
		1110
		1111	sub end {
		1112	return if --$_[0]{_ae_counter};
		1113	&{ $_[0]{_ae_end_cb} || sub { $_[0]->send } };
		1114	}
		1115
		1116	# undocumented/compatibility with pre-3.4
		1117	*broadcast = \&send;
		1118	*wait = \&_wait;
		1119
480	=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE	1120	=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE
		1121
		1122	This is an advanced topic that you do not normally need to use AnyEvent in
		1123	a module. This section is only of use to event loop authors who want to
		1124	provide AnyEvent compatibility.
481		1125
482	If you need to support another event library which isn't directly	1126	If you need to support another event library which isn't directly
483	supported by AnyEvent, you can supply your own interface to it by	1127	supported by AnyEvent, you can supply your own interface to it by
484	pushing, before the first watcher gets created, the package name of	1128	pushing, before the first watcher gets created, the package name of
485	the event module and the package name of the interface to use onto	1129	the event module and the package name of the interface to use onto
486	C<@AnyEvent::REGISTRY>. You can do that before and even without loading	1130	C<@AnyEvent::REGISTRY>. You can do that before and even without loading
487	AnyEvent.	1131	AnyEvent, so it is reasonably cheap.
488		1132
489	Example:	1133	Example:
490		1134
491	push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::];	1135	push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::];
492		1136
493	This tells AnyEvent to (literally) use the C<urxvt::anyevent::>	1137	This tells AnyEvent to (literally) use the C<urxvt::anyevent::>
494	package/class when it finds the C<urxvt> package/module is loaded. When	1138	package/class when it finds the C<urxvt> package/module is already loaded.
		1139
495	AnyEvent is loaded and asked to find a suitable event model, it will	1140	When AnyEvent is loaded and asked to find a suitable event model, it
496	first check for the presence of urxvt.	1141	will first check for the presence of urxvt by trying to C<use> the
		1142	C<urxvt::anyevent> module.
497		1143
498	The class should provide implementations for all watcher types (see	1144	The class should provide implementations for all watcher types. See
499	L<AnyEvent::Impl::Event> (source code), L<AnyEvent::Impl::Glib>	1145	L<AnyEvent::Impl::EV> (source code), L<AnyEvent::Impl::Glib> (Source code)
500	(Source code) and so on for actual examples, use C<perldoc -m	1146	and so on for actual examples. Use C<perldoc -m AnyEvent::Impl::Glib> to
501	AnyEvent::Impl::Glib> to see the sources).	1147	see the sources.
502		1148
		1149	If you don't provide C<signal> and C<child> watchers than AnyEvent will
		1150	provide suitable (hopefully) replacements.
		1151
503	The above isn't fictitious, the I<rxvt-unicode> (a.k.a. urxvt)	1152	The above example isn't fictitious, the I<rxvt-unicode> (a.k.a. urxvt)
504	uses the above line as-is. An interface isn't included in AnyEvent	1153	terminal emulator uses the above line as-is. An interface isn't included
505	because it doesn't make sense outside the embedded interpreter inside	1154	in AnyEvent because it doesn't make sense outside the embedded interpreter
506	I<rxvt-unicode>, and it is updated and maintained as part of the	1155	inside I<rxvt-unicode>, and it is updated and maintained as part of the
507	I<rxvt-unicode> distribution.	1156	I<rxvt-unicode> distribution.
508		1157
509	I<rxvt-unicode> also cheats a bit by not providing blocking access to	1158	I<rxvt-unicode> also cheats a bit by not providing blocking access to
510	condition variables: code blocking while waiting for a condition will	1159	condition variables: code blocking while waiting for a condition will
511	C<die>. This still works with most modules/usages, and blocking calls must	1160	C<die>. This still works with most modules/usages, and blocking calls must
512	not be in an interactive application, so it makes sense.	1161	not be done in an interactive application, so it makes sense.
513		1162
514	=head1 ENVIRONMENT VARIABLES	1163	=head1 ENVIRONMENT VARIABLES
515		1164
516	The following environment variables are used by this module:	1165	The following environment variables are used by this module:
517		1166
518	C<PERL_ANYEVENT_VERBOSE> when set to C<2> or higher, reports which event	1167	=over 4
519	model gets used.
520		1168
		1169	=item C<PERL_ANYEVENT_VERBOSE>
		1170
		1171	By default, AnyEvent will be completely silent except in fatal
		1172	conditions. You can set this environment variable to make AnyEvent more
		1173	talkative.
		1174
		1175	When set to C<1> or higher, causes AnyEvent to warn about unexpected
		1176	conditions, such as not being able to load the event model specified by
		1177	C<PERL_ANYEVENT_MODEL>.
		1178
		1179	When set to C<2> or higher, cause AnyEvent to report to STDERR which event
		1180	model it chooses.
		1181
		1182	=item C<PERL_ANYEVENT_MODEL>
		1183
		1184	This can be used to specify the event model to be used by AnyEvent, before
		1185	auto detection and -probing kicks in. It must be a string consisting
		1186	entirely of ASCII letters. The string C<AnyEvent::Impl::> gets prepended
		1187	and the resulting module name is loaded and if the load was successful,
		1188	used as event model. If it fails to load AnyEvent will proceed with
		1189	auto detection and -probing.
		1190
		1191	This functionality might change in future versions.
		1192
		1193	For example, to force the pure perl model (L<AnyEvent::Impl::Perl>) you
		1194	could start your program like this:
		1195
		1196	PERL_ANYEVENT_MODEL=Perl perl ...
		1197
		1198	=item C<PERL_ANYEVENT_PROTOCOLS>
		1199
		1200	Used by both L<AnyEvent::DNS> and L<AnyEvent::Socket> to determine preferences
		1201	for IPv4 or IPv6. The default is unspecified (and might change, or be the result
		1202	of auto probing).
		1203
		1204	Must be set to a comma-separated list of protocols or address families,
		1205	current supported: C<ipv4> and C<ipv6>. Only protocols mentioned will be
		1206	used, and preference will be given to protocols mentioned earlier in the
		1207	list.
		1208
		1209	This variable can effectively be used for denial-of-service attacks
		1210	against local programs (e.g. when setuid), although the impact is likely
		1211	small, as the program has to handle connection errors already-
		1212
		1213	Examples: C<PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6> - prefer IPv4 over IPv6,
		1214	but support both and try to use both. C<PERL_ANYEVENT_PROTOCOLS=ipv4>
		1215	- only support IPv4, never try to resolve or contact IPv6
		1216	addresses. C<PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4> support either IPv4 or
		1217	IPv6, but prefer IPv6 over IPv4.
		1218
		1219	=item C<PERL_ANYEVENT_EDNS0>
		1220
		1221	Used by L<AnyEvent::DNS> to decide whether to use the EDNS0 extension
		1222	for DNS. This extension is generally useful to reduce DNS traffic, but
		1223	some (broken) firewalls drop such DNS packets, which is why it is off by
		1224	default.
		1225
		1226	Setting this variable to C<1> will cause L<AnyEvent::DNS> to announce
		1227	EDNS0 in its DNS requests.
		1228
		1229	=item C<PERL_ANYEVENT_MAX_FORKS>
		1230
		1231	The maximum number of child processes that C<AnyEvent::Util::fork_call>
		1232	will create in parallel.
		1233
		1234	=back
		1235
521	=head1 EXAMPLE	1236	=head1 EXAMPLE PROGRAM
522		1237
523	The following program uses an io watcher to read data from stdin, a timer	1238	The following program uses an I/O watcher to read data from STDIN, a timer
524	to display a message once per second, and a condvar to exit the program	1239	to display a message once per second, and a condition variable to quit the
525	when the user enters quit:	1240	program when the user enters quit:
526		1241
527	use AnyEvent;	1242	use AnyEvent;
528		1243
529	my $cv = AnyEvent->condvar;	1244	my $cv = AnyEvent->condvar;
530		1245
531	my $io_watcher = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {	1246	my $io_watcher = AnyEvent->io (
		1247	fh => \*STDIN,
		1248	poll => 'r',
		1249	cb => sub {
532	warn "io event <$_[0]>\n"; # will always output <r>	1250	warn "io event <$_[0]>\n"; # will always output <r>
533	chomp (my $input = <STDIN>); # read a line	1251	chomp (my $input = <STDIN>); # read a line
534	warn "read: $input\n"; # output what has been read	1252	warn "read: $input\n"; # output what has been read
535	$cv->broadcast if $input =~ /^q/i; # quit program if /^q/i	1253	$cv->send if $input =~ /^q/i; # quit program if /^q/i
		1254	},
536	});	1255	);
537		1256
538	my $time_watcher; # can only be used once	1257	my $time_watcher; # can only be used once
539		1258
540	sub new_timer {	1259	sub new_timer {
541	$timer = AnyEvent->timer (after => 1, cb => sub {	1260	$timer = AnyEvent->timer (after => 1, cb => sub {
…		…
544	});	1263	});
545	}	1264	}
546		1265
547	new_timer; # create first timer	1266	new_timer; # create first timer
548		1267
549	$cv->wait; # wait until user enters /^q/i	1268	$cv->recv; # wait until user enters /^q/i
550		1269
551	=head1 REAL-WORLD EXAMPLE	1270	=head1 REAL-WORLD EXAMPLE
552		1271
553	Consider the L<Net::FCP> module. It features (among others) the following	1272	Consider the L<Net::FCP> module. It features (among others) the following
554	API calls, which are to freenet what HTTP GET requests are to http:	1273	API calls, which are to freenet what HTTP GET requests are to http:
…		…
604	syswrite $txn->{fh}, $txn->{request}	1323	syswrite $txn->{fh}, $txn->{request}
605	or die "connection or write error";	1324	or die "connection or write error";
606	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });	1325	$txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });
607		1326
608	Again, C<fh_ready_r> waits till all data has arrived, and then stores the	1327	Again, C<fh_ready_r> waits till all data has arrived, and then stores the
609	result and signals any possible waiters that the request ahs finished:	1328	result and signals any possible waiters that the request has finished:
610		1329
611	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};	1330	sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};
612		1331
613	if (end-of-file or data complete) {	1332	if (end-of-file or data complete) {
614	$txn->{result} = $txn->{buf};	1333	$txn->{result} = $txn->{buf};
615	$txn->{finished}->broadcast;	1334	$txn->{finished}->send;
616	$txb->{cb}->($txn) of $txn->{cb}; # also call callback	1335	$txb->{cb}->($txn) of $txn->{cb}; # also call callback
617	}	1336	}
618		1337
619	The C<result> method, finally, just waits for the finished signal (if the	1338	The C<result> method, finally, just waits for the finished signal (if the
620	request was already finished, it doesn't wait, of course, and returns the	1339	request was already finished, it doesn't wait, of course, and returns the
621	data:	1340	data:
622		1341
623	$txn->{finished}->wait;	1342	$txn->{finished}->recv;
624	return $txn->{result};	1343	return $txn->{result};
625		1344
626	The actual code goes further and collects all errors (C<die>s, exceptions)	1345	The actual code goes further and collects all errors (C<die>s, exceptions)
627	that occured during request processing. The C<result> method detects	1346	that occurred during request processing. The C<result> method detects
628	wether an exception as thrown (it is stored inside the $txn object)	1347	whether an exception as thrown (it is stored inside the $txn object)
629	and just throws the exception, which means connection errors and other	1348	and just throws the exception, which means connection errors and other
630	problems get reported tot he code that tries to use the result, not in a	1349	problems get reported tot he code that tries to use the result, not in a
631	random callback.	1350	random callback.
632		1351
633	All of this enables the following usage styles:	1352	All of this enables the following usage styles:
634		1353
635	1. Blocking:	1354	1. Blocking:
636		1355
637	my $data = $fcp->client_get ($url);	1356	my $data = $fcp->client_get ($url);
638		1357
639	2. Blocking, but parallelizing:	1358	2. Blocking, but running in parallel:
640		1359
641	my @datas = map $_->result,	1360	my @datas = map $_->result,
642	map $fcp->txn_client_get ($_),	1361	map $fcp->txn_client_get ($_),
643	@urls;	1362	@urls;
644		1363
645	Both blocking examples work without the module user having to know	1364	Both blocking examples work without the module user having to know
646	anything about events.	1365	anything about events.
647		1366
648	3a. Event-based in a main program, using any support Event module:	1367	3a. Event-based in a main program, using any supported event module:
649		1368
650	use Event;	1369	use EV;
651		1370
652	$fcp->txn_client_get ($url)->cb (sub {	1371	$fcp->txn_client_get ($url)->cb (sub {
653	my $txn = shift;	1372	my $txn = shift;
654	my $data = $txn->result;	1373	my $data = $txn->result;
655	...	1374	...
656	});	1375	});
657		1376
658	Event::loop;	1377	EV::loop;
659		1378
660	3b. The module user could use AnyEvent, too:	1379	3b. The module user could use AnyEvent, too:
661		1380
662	use AnyEvent;	1381	use AnyEvent;
663		1382
664	my $quit = AnyEvent->condvar;	1383	my $quit = AnyEvent->condvar;
665		1384
666	$fcp->txn_client_get ($url)->cb (sub {	1385	$fcp->txn_client_get ($url)->cb (sub {
667	...	1386	...
668	$quit->broadcast;	1387	$quit->send;
669	});	1388	});
670		1389
671	$quit->wait;	1390	$quit->recv;
		1391
		1392
		1393	=head1 BENCHMARKS
		1394
		1395	To give you an idea of the performance and overheads that AnyEvent adds
		1396	over the event loops themselves and to give you an impression of the speed
		1397	of various event loops I prepared some benchmarks.
		1398
		1399	=head2 BENCHMARKING ANYEVENT OVERHEAD
		1400
		1401	Here is a benchmark of various supported event models used natively and
		1402	through AnyEvent. The benchmark creates a lot of timers (with a zero
		1403	timeout) and I/O watchers (watching STDOUT, a pty, to become writable,
		1404	which it is), lets them fire exactly once and destroys them again.
		1405
		1406	Source code for this benchmark is found as F<eg/bench> in the AnyEvent
		1407	distribution.
		1408
		1409	=head3 Explanation of the columns
		1410
		1411	I<watcher> is the number of event watchers created/destroyed. Since
		1412	different event models feature vastly different performances, each event
		1413	loop was given a number of watchers so that overall runtime is acceptable
		1414	and similar between tested event loop (and keep them from crashing): Glib
		1415	would probably take thousands of years if asked to process the same number
		1416	of watchers as EV in this benchmark.
		1417
		1418	I<bytes> is the number of bytes (as measured by the resident set size,
		1419	RSS) consumed by each watcher. This method of measuring captures both C
		1420	and Perl-based overheads.
		1421
		1422	I<create> is the time, in microseconds (millionths of seconds), that it
		1423	takes to create a single watcher. The callback is a closure shared between
		1424	all watchers, to avoid adding memory overhead. That means closure creation
		1425	and memory usage is not included in the figures.
		1426
		1427	I<invoke> is the time, in microseconds, used to invoke a simple
		1428	callback. The callback simply counts down a Perl variable and after it was
		1429	invoked "watcher" times, it would C<< ->send >> a condvar once to
		1430	signal the end of this phase.
		1431
		1432	I<destroy> is the time, in microseconds, that it takes to destroy a single
		1433	watcher.
		1434
		1435	=head3 Results
		1436
		1437	name watchers bytes create invoke destroy comment
		1438	EV/EV 400000 244 0.56 0.46 0.31 EV native interface
		1439	EV/Any 100000 244 2.50 0.46 0.29 EV + AnyEvent watchers
		1440	CoroEV/Any 100000 244 2.49 0.44 0.29 coroutines + Coro::Signal
		1441	Perl/Any 100000 513 4.92 0.87 1.12 pure perl implementation
		1442	Event/Event 16000 516 31.88 31.30 0.85 Event native interface
		1443	Event/Any 16000 590 35.75 31.42 1.08 Event + AnyEvent watchers
		1444	Glib/Any 16000 1357 98.22 12.41 54.00 quadratic behaviour
		1445	Tk/Any 2000 1860 26.97 67.98 14.00 SEGV with >> 2000 watchers
		1446	POE/Event 2000 6644 108.64 736.02 14.73 via POE::Loop::Event
		1447	POE/Select 2000 6343 94.13 809.12 565.96 via POE::Loop::Select
		1448
		1449	=head3 Discussion
		1450
		1451	The benchmark does I<not> measure scalability of the event loop very
		1452	well. For example, a select-based event loop (such as the pure perl one)
		1453	can never compete with an event loop that uses epoll when the number of
		1454	file descriptors grows high. In this benchmark, all events become ready at
		1455	the same time, so select/poll-based implementations get an unnatural speed
		1456	boost.
		1457
		1458	Also, note that the number of watchers usually has a nonlinear effect on
		1459	overall speed, that is, creating twice as many watchers doesn't take twice
		1460	the time - usually it takes longer. This puts event loops tested with a
		1461	higher number of watchers at a disadvantage.
		1462
		1463	To put the range of results into perspective, consider that on the
		1464	benchmark machine, handling an event takes roughly 1600 CPU cycles with
		1465	EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000 CPU
		1466	cycles with POE.
		1467
		1468	C<EV> is the sole leader regarding speed and memory use, which are both
		1469	maximal/minimal, respectively. Even when going through AnyEvent, it uses
		1470	far less memory than any other event loop and is still faster than Event
		1471	natively.
		1472
		1473	The pure perl implementation is hit in a few sweet spots (both the
		1474	constant timeout and the use of a single fd hit optimisations in the perl
		1475	interpreter and the backend itself). Nevertheless this shows that it
		1476	adds very little overhead in itself. Like any select-based backend its
		1477	performance becomes really bad with lots of file descriptors (and few of
		1478	them active), of course, but this was not subject of this benchmark.
		1479
		1480	The C<Event> module has a relatively high setup and callback invocation
		1481	cost, but overall scores in on the third place.
		1482
		1483	C<Glib>'s memory usage is quite a bit higher, but it features a
		1484	faster callback invocation and overall ends up in the same class as
		1485	C<Event>. However, Glib scales extremely badly, doubling the number of
		1486	watchers increases the processing time by more than a factor of four,
		1487	making it completely unusable when using larger numbers of watchers
		1488	(note that only a single file descriptor was used in the benchmark, so
		1489	inefficiencies of C<poll> do not account for this).
		1490
		1491	The C<Tk> adaptor works relatively well. The fact that it crashes with
		1492	more than 2000 watchers is a big setback, however, as correctness takes
		1493	precedence over speed. Nevertheless, its performance is surprising, as the
		1494	file descriptor is dup()ed for each watcher. This shows that the dup()
		1495	employed by some adaptors is not a big performance issue (it does incur a
		1496	hidden memory cost inside the kernel which is not reflected in the figures
		1497	above).
		1498
		1499	C<POE>, regardless of underlying event loop (whether using its pure perl
		1500	select-based backend or the Event module, the POE-EV backend couldn't
		1501	be tested because it wasn't working) shows abysmal performance and
		1502	memory usage with AnyEvent: Watchers use almost 30 times as much memory
		1503	as EV watchers, and 10 times as much memory as Event (the high memory
		1504	requirements are caused by requiring a session for each watcher). Watcher
		1505	invocation speed is almost 900 times slower than with AnyEvent's pure perl
		1506	implementation.
		1507
		1508	The design of the POE adaptor class in AnyEvent can not really account
		1509	for the performance issues, though, as session creation overhead is
		1510	small compared to execution of the state machine, which is coded pretty
		1511	optimally within L<AnyEvent::Impl::POE> (and while everybody agrees that
		1512	using multiple sessions is not a good approach, especially regarding
		1513	memory usage, even the author of POE could not come up with a faster
		1514	design).
		1515
		1516	=head3 Summary
		1517
		1518	=over 4
		1519
		1520	=item * Using EV through AnyEvent is faster than any other event loop
		1521	(even when used without AnyEvent), but most event loops have acceptable
		1522	performance with or without AnyEvent.
		1523
		1524	=item * The overhead AnyEvent adds is usually much smaller than the overhead of
		1525	the actual event loop, only with extremely fast event loops such as EV
		1526	adds AnyEvent significant overhead.
		1527
		1528	=item * You should avoid POE like the plague if you want performance or
		1529	reasonable memory usage.
		1530
		1531	=back
		1532
		1533	=head2 BENCHMARKING THE LARGE SERVER CASE
		1534
		1535	This benchmark actually benchmarks the event loop itself. It works by
		1536	creating a number of "servers": each server consists of a socket pair, a
		1537	timeout watcher that gets reset on activity (but never fires), and an I/O
		1538	watcher waiting for input on one side of the socket. Each time the socket
		1539	watcher reads a byte it will write that byte to a random other "server".
		1540
		1541	The effect is that there will be a lot of I/O watchers, only part of which
		1542	are active at any one point (so there is a constant number of active
		1543	fds for each loop iteration, but which fds these are is random). The
		1544	timeout is reset each time something is read because that reflects how
		1545	most timeouts work (and puts extra pressure on the event loops).
		1546
		1547	In this benchmark, we use 10000 socket pairs (20000 sockets), of which 100
		1548	(1%) are active. This mirrors the activity of large servers with many
		1549	connections, most of which are idle at any one point in time.
		1550
		1551	Source code for this benchmark is found as F<eg/bench2> in the AnyEvent
		1552	distribution.
		1553
		1554	=head3 Explanation of the columns
		1555
		1556	I<sockets> is the number of sockets, and twice the number of "servers" (as
		1557	each server has a read and write socket end).
		1558
		1559	I<create> is the time it takes to create a socket pair (which is
		1560	nontrivial) and two watchers: an I/O watcher and a timeout watcher.
		1561
		1562	I<request>, the most important value, is the time it takes to handle a
		1563	single "request", that is, reading the token from the pipe and forwarding
		1564	it to another server. This includes deleting the old timeout and creating
		1565	a new one that moves the timeout into the future.
		1566
		1567	=head3 Results
		1568
		1569	name sockets create request
		1570	EV 20000 69.01 11.16
		1571	Perl 20000 73.32 35.87
		1572	Event 20000 212.62 257.32
		1573	Glib 20000 651.16 1896.30
		1574	POE 20000 349.67 12317.24 uses POE::Loop::Event
		1575
		1576	=head3 Discussion
		1577
		1578	This benchmark I<does> measure scalability and overall performance of the
		1579	particular event loop.
		1580
		1581	EV is again fastest. Since it is using epoll on my system, the setup time
		1582	is relatively high, though.
		1583
		1584	Perl surprisingly comes second. It is much faster than the C-based event
		1585	loops Event and Glib.
		1586
		1587	Event suffers from high setup time as well (look at its code and you will
		1588	understand why). Callback invocation also has a high overhead compared to
		1589	the C<< $_->() for .. >>-style loop that the Perl event loop uses. Event
		1590	uses select or poll in basically all documented configurations.
		1591
		1592	Glib is hit hard by its quadratic behaviour w.r.t. many watchers. It
		1593	clearly fails to perform with many filehandles or in busy servers.
		1594
		1595	POE is still completely out of the picture, taking over 1000 times as long
		1596	as EV, and over 100 times as long as the Perl implementation, even though
		1597	it uses a C-based event loop in this case.
		1598
		1599	=head3 Summary
		1600
		1601	=over 4
		1602
		1603	=item * The pure perl implementation performs extremely well.
		1604
		1605	=item * Avoid Glib or POE in large projects where performance matters.
		1606
		1607	=back
		1608
		1609	=head2 BENCHMARKING SMALL SERVERS
		1610
		1611	While event loops should scale (and select-based ones do not...) even to
		1612	large servers, most programs we (or I :) actually write have only a few
		1613	I/O watchers.
		1614
		1615	In this benchmark, I use the same benchmark program as in the large server
		1616	case, but it uses only eight "servers", of which three are active at any
		1617	one time. This should reflect performance for a small server relatively
		1618	well.
		1619
		1620	The columns are identical to the previous table.
		1621
		1622	=head3 Results
		1623
		1624	name sockets create request
		1625	EV 16 20.00 6.54
		1626	Perl 16 25.75 12.62
		1627	Event 16 81.27 35.86
		1628	Glib 16 32.63 15.48
		1629	POE 16 261.87 276.28 uses POE::Loop::Event
		1630
		1631	=head3 Discussion
		1632
		1633	The benchmark tries to test the performance of a typical small
		1634	server. While knowing how various event loops perform is interesting, keep
		1635	in mind that their overhead in this case is usually not as important, due
		1636	to the small absolute number of watchers (that is, you need efficiency and
		1637	speed most when you have lots of watchers, not when you only have a few of
		1638	them).
		1639
		1640	EV is again fastest.
		1641
		1642	Perl again comes second. It is noticeably faster than the C-based event
		1643	loops Event and Glib, although the difference is too small to really
		1644	matter.
		1645
		1646	POE also performs much better in this case, but is is still far behind the
		1647	others.
		1648
		1649	=head3 Summary
		1650
		1651	=over 4
		1652
		1653	=item * C-based event loops perform very well with small number of
		1654	watchers, as the management overhead dominates.
		1655
		1656	=back
		1657
		1658
		1659	=head1 FORK
		1660
		1661	Most event libraries are not fork-safe. The ones who are usually are
		1662	because they rely on inefficient but fork-safe C<select> or C<poll>
		1663	calls. Only L<EV> is fully fork-aware.
		1664
		1665	If you have to fork, you must either do so I<before> creating your first
		1666	watcher OR you must not use AnyEvent at all in the child.
		1667
		1668
		1669	=head1 SECURITY CONSIDERATIONS
		1670
		1671	AnyEvent can be forced to load any event model via
		1672	$ENV{PERL_ANYEVENT_MODEL}. While this cannot (to my knowledge) be used to
		1673	execute arbitrary code or directly gain access, it can easily be used to
		1674	make the program hang or malfunction in subtle ways, as AnyEvent watchers
		1675	will not be active when the program uses a different event model than
		1676	specified in the variable.
		1677
		1678	You can make AnyEvent completely ignore this variable by deleting it
		1679	before the first watcher gets created, e.g. with a C<BEGIN> block:
		1680
		1681	BEGIN { delete $ENV{PERL_ANYEVENT_MODEL} }
		1682
		1683	use AnyEvent;
		1684
		1685	Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can
		1686	be used to probe what backend is used and gain other information (which is
		1687	probably even less useful to an attacker than PERL_ANYEVENT_MODEL).
		1688
		1689
		1690	=head1 BUGS
		1691
		1692	Perl 5.8 has numerous memleaks that sometimes hit this module and are hard
		1693	to work around. If you suffer from memleaks, first upgrade to Perl 5.10
		1694	and check wether the leaks still show up. (Perl 5.10.0 has other annoying
		1695	mamleaks, such as leaking on C<map> and C<grep> but it is usually not as
		1696	pronounced).
		1697
672		1698
673	=head1 SEE ALSO	1699	=head1 SEE ALSO
674		1700
675	Event modules: L<Coro::Event>, L<Coro>, L<Event>, L<Glib::Event>, L<Glib>.	1701	Utility functions: L<AnyEvent::Util>.
676		1702
677	Implementations: L<AnyEvent::Impl::Coro>, L<AnyEvent::Impl::Event>, L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>.	1703	Event modules: L<EV>, L<EV::Glib>, L<Glib::EV>, L<Event>, L<Glib::Event>,
		1704	L<Glib>, L<Tk>, L<Event::Lib>, L<Qt>, L<POE>.
678		1705
679	Nontrivial usage example: L<Net::FCP>.	1706	Implementations: L<AnyEvent::Impl::EV>, L<AnyEvent::Impl::Event>,
		1707	L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>, L<AnyEvent::Impl::Perl>,
		1708	L<AnyEvent::Impl::EventLib>, L<AnyEvent::Impl::Qt>,
		1709	L<AnyEvent::Impl::POE>.
680		1710
681	=head1	1711	Non-blocking file handles, sockets, TCP clients and
		1712	servers: L<AnyEvent::Handle>, L<AnyEvent::Socket>.
		1713
		1714	Asynchronous DNS: L<AnyEvent::DNS>.
		1715
		1716	Coroutine support: L<Coro>, L<Coro::AnyEvent>, L<Coro::EV>, L<Coro::Event>,
		1717
		1718	Nontrivial usage examples: L<Net::FCP>, L<Net::XMPP2>, L<AnyEvent::DNS>.
		1719
		1720
		1721	=head1 AUTHOR
		1722
		1723	Marc Lehmann <schmorp@schmorp.de>
		1724	http://home.schmorp.de/
682		1725
683	=cut	1726	=cut
684		1727
685	1	1728	1
686		1729

Diff Legend

–	Removed lines
+	Added lines
<	Changed lines
>	Changed lines